Nvidia is facing a new test of the artificial intelligence investment cycle as rapidly increasing demand for compute power coincides with accelerating revenue at leading AI companies. An analysis endorsed by investor Gavin Baker suggests Anthropic’s economics have improved sharply, reinforcing the argument that cheaper AI tokens may ultimately drive greater overall spending on compute and strengthen demand for Nvidia’s graphics processing units.

Anthropic’s Revenue Growth Changes the Compute Equation

Anthropic’s S-1 filing provides a notable indicator of how quickly the economics of generative AI are changing. According to the analysis cited in the source, Anthropic generated $2.08 of revenue for every dollar of compute, compared with only $0.40 a year earlier. The sharp increase suggests that improving monetization and efficiency can allow AI developers to generate substantially more revenue from the computing resources they deploy.

For Nvidia, the significance extends beyond Anthropic itself. If AI companies can reduce the cost of generating tokens while simultaneously expanding usage, lower prices could stimulate greater consumption. Baker’s argument is that cheaper tokens do not necessarily mean lower aggregate spending; instead, they could encourage broader adoption and increase the total amount of computing capacity required.

Why Lower AI Costs Could Support Nvidia

The dynamic creates an important feedback loop for the AI semiconductor market. Lower inference costs can make AI applications more accessible, potentially increasing demand for services and encouraging developers and enterprises to run more workloads. As utilization expands, AI companies may require additional GPUs and related infrastructure, supporting Nvidia’s position as a major supplier to the sector.

That relationship is particularly important because Nvidia’s growth expectations have already moved to exceptionally high levels. Analysts cited in the source expect Nvidia revenue to increase from $215.9 billion in fiscal 2026 to $692 billion in fiscal 2028. Such forecasts imply that investors are pricing in continued expansion of AI infrastructure rather than a short-lived technology spending cycle.

Growth Comes With Margin Pressure

Strong demand does not eliminate the financial pressures associated with rapid expansion. Rising memory costs are expected to reduce Nvidia’s gross margin to between 71% and 72% in the January 2027 quarter. For a company operating at the center of the AI hardware boom, maintaining margins while component costs increase will be an important measure of pricing power and supply-chain management.

Valuation also remains a key consideration. Nvidia is trading at approximately 18 times forward earnings, compared with a three-year average of 30 times, according to the source. That relative valuation could attract attention as investors assess whether the company’s earnings growth can continue at a pace that supports its current market expectations.

November Earnings Will Test the AI Spending Thesis

Nvidia’s November 17 earnings report will provide an important test of the market’s assumptions. Investors will be watching revenue growth, margins, demand indicators and management’s outlook for AI infrastructure spending. The broader question is whether improving economics at AI developers such as Anthropic will translate into a sustained increase in compute consumption.

If falling token costs lead to significantly greater AI usage, Nvidia could continue benefiting from a cycle in which efficiency creates more demand rather than less. At the same time, rising component costs, elevated expectations and the scale of projected revenue growth mean that even strong results could be judged against an increasingly demanding benchmark.